Skip to contents

Conditional Effects

Source: R/condeff.R, R/condeff_boot.R
cond_effect.Rd

Compute the conditional effects in a moderated regression model.

Usage

cond_effect(
  output,
  x = NULL,
  w = NULL,
  w_method = c("sd", "percentile"),
  w_percentiles = c(0.16, 0.5, 0.84),
  w_sd_to_percentiles = NA,
  w_from_mean_in_sd = 1,
  w_values = NULL
)

cond_effect_boot(
  output,
  x = NULL,
  w = NULL,
  ...,
  conf = 0.95,
  nboot = 100,
  boot_args = NULL,
  save_boot_est = TRUE,
  full_output = FALSE,
  do_boot = TRUE
)

Arguments

output

The output from stats::lm(). It can also accept the output from std_selected() or std_selected_boot().

x

The focal variable (independent variable), that is, the variable with its effect on the outcome variable (dependent) being moderated. It must be a numeric variable.

w

The moderator. Both numeric variables and categorical variables (character or factor) are supported.

w_method

How to define "low", "medium", and "high" for the moderator levels. Default is in terms of mean and standard deviation (SD) of the moderator, "sd": "low", "medium", and "high" are one SD below mean, mean, and one SD above mean, respectively. If equal to "percentile", then percentiles of the moderator in the dataset are used: "low", "medium", and "high" are 16th, 50th (median), and 84th percentiles, respectively. Ignored if w is categorical.

w_percentiles

If w_method is "percentile", then this argument specifies the three percentiles to be used, divided by 100. It must be a vector of two numbers. The default is c(.16, .50, .84), the 16th, 50th, and 84th percentiles, which corresponds approximately to one SD below and above mean in a normal distribution, respectively. Ignored if w is categorical.

w_sd_to_percentiles

If w_method is "percentile" and this argument is set to a number, this number will be used to to determine the percentiles to be used. The lower percentile is the percentile in a normal distribution that is w_sd_to_percentiles SD below the mean. The upper percentile is the percentile in a normal distribution that is w_sd_to_percentiles SD above the mean. Therefore, if w_sd_to_percentiles is set to 1, then the lower and upper percentiles are 16th and 84th, respectively. Default is NA.

w_from_mean_in_sd

How many SD from mean is used to define "low" and "high" for the moderator. Default is 1. Ignored if w is categorical.

w_values

The values of w to be used. Default is NULL. If a numeric vector is supplied, these values will be used to compute the conditional effects. Other arguments on generating levels are ignored. Note that, if w has been standardized or centered, these values are for the standardized or centered w. The values will always be sorted. This argument is ignored if w is categorical.

...

Arguments to be passed to cond_effect().

conf

The level of confidence for the confidence interval. Default is .95, to get 95% confidence intervals.

nboot

The number of bootstrap samples. Default is 100.

boot_args

A named list of arguments to be passed to boot::boot(). Default is NULL.

save_boot_est

If TRUE, the default, the bootstrap estimates will be saved in the element boot_est of the output.

full_output

Whether the full output from boot::boot() will be returned. Default is FALSE. If TRUE, the full output from boot::boot() will be saved in the element boot_out of the output.

do_boot

Whether bootstrapping confidence intervals will be formed. Default is TRUE. If FALSE, all arguments related to bootstrapping will be ignored.

Value

cond_effect() returns a data-frame-like object of the conditional effects. The class is cond_effect and the print method will print additional information of the conditional effects. Additional information is stored in the following attributes:

  • call: The original call.

  • output: The output object, such as the output from lm().

  • x, y, and w: The three variables used to compute the conditional effects: focal variable (x), outcome variable (y), and the moderator (w).

  • w_method: The method used to determine the values of the moderator at the selected levels.

  • w_percentiles The percentiles to use if w_method = "percentile".

  • w_sd_to_percentiles: If not equal to NA, this is a scalar, the number of standard deviation from the mean used to determine the percentiles for the "low" and "high" levels of the moderator.

  • w_from_mean_in_sd: The number of SD above or below the mean, for determining the "low" and "high" levels of the moderator if w_method is "sd".

  • w_empirical_percentiles: The actual percentile levels in the dataset for the selected levels of the moderator. A numeric vector.

  • w_empirical_z: The actual distance from the mean, in SD, of each selected level of the moderator. A numeric vector.

  • y_standardized, x_standardized, and w_standardized: Each of them is a logical scalar, indicating whether the outcome variable, focal variable, and moderator are standardized.

cond_effect_boot() also returns a data-frame-like object of the conditional effects of the class cond_effect, with additional information from the bootstrapping stored in these attributes:

  • boot_ci: A data frame of the bootstrap confidence intervals of the conditional effects.

  • nboot: The number of bootstrap samples requested.

  • conf: The level of confidence, in proportion.

  • boot_est: A matrix of the bootstrap estimates of the conditional effects. The number of rows equal to nboot, and the number of columns equal to the number of levels of the moderator.

  • cond_effect_boot_call: The call to cond_effect_boot().

  • boot_out: If available, the original output from boot::boot().

Details

cond_effect() uses the centering approach to find the conditional effect of the focal variable. For each level of the moderator, the value for this level is subtracted from the moderator scores, and the model is fitted to the modified data. The coefficient of the focal variable is then the conditional effect of the focal variable when the moderator's score is equal this value.

cond_effect_boot() function is a wrapper of cond_effect(). It calls cond_effect() once for each bootstrap sample, and then computes the nonparametric bootstrap percentile confidence intervals (Cheung, Cheung, Lau, Hui, & Vong, 2022). If the output object is the output of std_selected() or std_selected_boot(), in which mean-centering and/or standardization have been conducted, they will be repeated in each bootstrap sample. Therefore, like std_selected_boot(), it can be used for form nonparametric bootstrap confidence intervals for standardized effects, though cond_effect_boot() does this for the standardized conditional effects.

This function ignores bootstrapping done by std_selected_boot(). It will do its own bootstrapping.

If do_boot is FALSE, then the object it returns is identical to that by cond_effect().

This function intentionally does not have an argument for setting the seed for random number. Users are recommended to set the seed, e.g., using set.seed() before calling it, to ensure reproducibility.

Functions

  • cond_effect_boot(): A wrapper of cond_effect() that forms nonparametric bootstrap confidence intervals.

Author

Shu Fai Cheung https://orcid.org/0000-0002-9871-9448

Examples


# Load a sample data set

dat <- test_x_1_w_1_v_1_cat1_n_500

# Do a moderated regression by lm
lm_raw <- lm(dv ~ iv*mod + v1 + cat1, dat)
summary(lm_raw)
#> 
#> Call:
#> lm(formula = dv ~ iv * mod + v1 + cat1, data = dat)
#> 
#> Residuals:
#>     Min      1Q  Median      3Q     Max 
#> -2146.0  -431.9   -25.0   411.2  2309.3 
#> 
#> Coefficients:
#>             Estimate Std. Error t value Pr(>|t|)  
#> (Intercept)  308.767   4075.066   0.076   0.9396  
#> iv            52.760    271.242   0.195   0.8459  
#> mod            5.127     40.772   0.126   0.9000  
#> v1           -12.760     10.174  -1.254   0.2104  
#> cat1gp2     -158.673     71.834  -2.209   0.0276 *
#> cat1gp3      -43.166     75.283  -0.573   0.5666  
#> iv:mod         3.416      2.709   1.261   0.2080  
#> ---
#> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#> 
#> Residual standard error: 665 on 493 degrees of freedom
#> Multiple R-squared:  0.6352,	Adjusted R-squared:  0.6307 
#> F-statistic:   143 on 6 and 493 DF,  p-value: < 2.2e-16
#> 
cond_effect(lm_raw, x = iv, w = mod)
#> The effects of iv on dv, conditional on mod:
#> 
#>   Level     mod iv Effect   S.E.      t     p Sig
#>    High 105.436   412.911 20.827 19.826 0.000 ***
#>  Medium 100.395   395.693 14.684 26.948 0.000 ***
#>     Low  95.354   378.474 19.249 19.662 0.000 ***
#> 
#> 
#> The regression model:
#> 
#> 	dv ~ iv * mod + v1 + cat1
#> 
#> Interpreting the levels of mod:
#> 
#>   Level     mod % Below From Mean (in SD)
#>    High 105.436   84.00              1.00
#>  Medium 100.395   47.40              0.00
#>     Low  95.354   17.20             -1.00
#> 
#> - % Below: The percent of cases equal to or less than a level.
#> - From Mean (in SD): Distance of a level from the mean, in standard
#>   deviation (+ve above, -ve below).

lm_std <- std_selected(lm_raw, to_standardize = ~ iv + mod)
cond_effect(lm_std, x = iv, w = mod)
#> The effects of iv on dv, conditional on mod:
#> 
#>   Level    mod iv Effect   S.E.      t     p Sig
#>    High  1.000   841.990 42.468 19.826 0.000 ***
#>  Medium  0.000   806.878 29.942 26.948 0.000 ***
#>     Low -1.000   771.767 39.251 19.662 0.000 ***
#> 
#> 
#> The regression model:
#> 
#> 	dv ~ iv * mod + v1 + cat1
#> 
#> Interpreting the levels of mod:
#> 
#>   Level    mod % Below From Mean (in SD)
#>    High  1.000   84.00              1.00
#>  Medium  0.000   47.40              0.00
#>     Low -1.000   17.20             -1.00
#> 
#> - % Below: The percent of cases equal to or less than a level.
#> - From Mean (in SD): Distance of a level from the mean, in standard
#>   deviation (+ve above, -ve below).
#> 
#> Note:
#> 
#> - The variable(s) iv, mod is/are standardized.
#> - One or more variables are scaled by SD or standardized. OLS standard
#>   errors and confidence intervals may be biased for their coefficients.
#>   Please use `cond_effect_boot()`.

# Categorical moderator
lm_cat <- lm(dv ~ iv*cat1 + v1, dat)
summary(lm_cat)
#> 
#> Call:
#> lm(formula = dv ~ iv * cat1 + v1, data = dat)
#> 
#> Residuals:
#>      Min       1Q   Median       3Q      Max 
#> -2457.67  -506.03     3.46   437.95  2738.18 
#> 
#> Coefficients:
#>             Estimate Std. Error t value Pr(>|t|)    
#> (Intercept)   979.20     459.19   2.132   0.0335 *  
#> iv            391.03      29.25  13.370   <2e-16 ***
#> cat1gp2      -845.49     584.85  -1.446   0.1489    
#> cat1gp3       259.55     620.76   0.418   0.6760    
#> v1            -19.36      11.00  -1.759   0.0791 .  
#> iv:cat1gp2     43.28      38.44   1.126   0.2608    
#> iv:cat1gp3    -21.22      41.08  -0.516   0.6058    
#> ---
#> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#> 
#> Residual standard error: 721.3 on 493 degrees of freedom
#> Multiple R-squared:  0.5707,	Adjusted R-squared:  0.5655 
#> F-statistic: 109.2 on 6 and 493 DF,  p-value: < 2.2e-16
#> 
cond_effect(lm_cat, x = iv, w = cat1)
#> The effects of iv on dv, conditional on cat1:
#> 
#>  Level cat1 iv Effect   S.E.      t     p Sig
#>    gp1  gp1   391.026 29.246 13.370 0.000 ***
#>    gp2  gp2   434.302 24.937 17.416 0.000 ***
#>    gp3  gp3   369.807 28.858 12.815 0.000 ***
#> 
#> 
#> The regression model:
#> 
#> 	dv ~ iv * cat1 + v1


# Load a sample data set

dat <- test_x_1_w_1_v_1_cat1_n_500

# Do a moderated regression by lm
lm_raw <- lm(dv ~ iv*mod + v1 + cat1, dat)
summary(lm_raw)
#> 
#> Call:
#> lm(formula = dv ~ iv * mod + v1 + cat1, data = dat)
#> 
#> Residuals:
#>     Min      1Q  Median      3Q     Max 
#> -2146.0  -431.9   -25.0   411.2  2309.3 
#> 
#> Coefficients:
#>             Estimate Std. Error t value Pr(>|t|)  
#> (Intercept)  308.767   4075.066   0.076   0.9396  
#> iv            52.760    271.242   0.195   0.8459  
#> mod            5.127     40.772   0.126   0.9000  
#> v1           -12.760     10.174  -1.254   0.2104  
#> cat1gp2     -158.673     71.834  -2.209   0.0276 *
#> cat1gp3      -43.166     75.283  -0.573   0.5666  
#> iv:mod         3.416      2.709   1.261   0.2080  
#> ---
#> Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
#> 
#> Residual standard error: 665 on 493 degrees of freedom
#> Multiple R-squared:  0.6352,	Adjusted R-squared:  0.6307 
#> F-statistic:   143 on 6 and 493 DF,  p-value: < 2.2e-16
#> 

lm_std <- std_selected(lm_raw, to_standardize = ~ iv + mod)
cond_effect(lm_std, x = iv, w = mod)
#> The effects of iv on dv, conditional on mod:
#> 
#>   Level    mod iv Effect   S.E.      t     p Sig
#>    High  1.000   841.990 42.468 19.826 0.000 ***
#>  Medium  0.000   806.878 29.942 26.948 0.000 ***
#>     Low -1.000   771.767 39.251 19.662 0.000 ***
#> 
#> 
#> The regression model:
#> 
#> 	dv ~ iv * mod + v1 + cat1
#> 
#> Interpreting the levels of mod:
#> 
#>   Level    mod % Below From Mean (in SD)
#>    High  1.000   84.00              1.00
#>  Medium  0.000   47.40              0.00
#>     Low -1.000   17.20             -1.00
#> 
#> - % Below: The percent of cases equal to or less than a level.
#> - From Mean (in SD): Distance of a level from the mean, in standard
#>   deviation (+ve above, -ve below).
#> 
#> Note:
#> 
#> - The variable(s) iv, mod is/are standardized.
#> - One or more variables are scaled by SD or standardized. OLS standard
#>   errors and confidence intervals may be biased for their coefficients.
#>   Please use `cond_effect_boot()`.

# Form nonparametric bootstrap confidence intervals
# Use 2000 or even 5000 for nboot in real research
out <- cond_effect_boot(lm_std, x = iv, w = mod, nboot = 50)
out
#> The effects of iv on dv, conditional on mod:
#> 
#>   Level    mod iv Effect CI Lower CI Upper   S.E.      t     p Sig
#>    High  1.000   841.990  745.556  969.398 42.468 19.826 0.000 ***
#>  Medium  0.000   806.878  731.389  882.734 29.942 26.948 0.000 ***
#>     Low -1.000   771.767  689.364  835.260 39.251 19.662 0.000 ***
#> 
#> [CI Lower, CI Upper] shows the 95% nonparametric bootstrap confidence
#> interval(s) (based on 50 bootstrap samples).
#> 
#> 
#> The regression model:
#> 
#> 	dv ~ iv * mod + v1 + cat1
#> 
#> Interpreting the levels of mod:
#> 
#>   Level    mod % Below From Mean (in SD)
#>    High  1.000   84.00              1.00
#>  Medium  0.000   47.40              0.00
#>     Low -1.000   17.20             -1.00
#> 
#> - % Below: The percent of cases equal to or less than a level.
#> - From Mean (in SD): Distance of a level from the mean, in standard
#>   deviation (+ve above, -ve below).
#> 
#> Note:
#> 
#> - The variable(s) iv, mod is/are standardized.